Glossarion
Glossarion is a GUI utility for AI-assisted novel translation using OpenAI and Gemini APIs. It supports contextual translation, glossary generation, and EPUB rebuilding with full control.
Glossarion translates novels and media with AI through a desktop interface
A PySide6 application turns EPUB, text, and subtitle files into context aware translations using many AI providers.
What Glossarion translates
Glossarion is a desktop translation application built with PySide6. The README describes it as a comprehensive AI powered translation suite for light novels, web novels, manga, subtitles, synchronized lyrics, and documents in virtually any language. It ships preconfigured profiles for Korean, Japanese, and Chinese through its profile system. The stated aim is to turn EPUB, TXT, PDF, SRT, ASS, and LRC files into high quality, context aware translations. The project is licensed under MIT and the primary language is Python, with Python 3.10 or newer required. The README links to a complete user guide dated August 21, 2026, which explains every button and setting in plain language, including how to translate for free. By centering the tool on a modern GUI, Glossarion gives the user control over each step instead of hiding the process behind a single button. The README presents the app as a suite rather than a one shot converter, with separate panels for different media and a glossary feature to keep names consistent across a long work. The profile system means language specific setup is mostly a selection rather than manual configuration. The information above is drawn from the project README and its repository metadata, which remain the source for the project's current behaviour. The app is built with PySide6, requires Python 3.10 or newer, and is licensed under MIT. A reader who wants exact commands or file names should open the repository and read the documentation the maintainers provide.
Supported file types and providers
The README lists the file types Glossarion handles: EPUB, TXT, PDF, SRT, ASS, and LRC. It also describes a dedicated manga translator panel that combines multiple AI models to translate .cbz archives and direct image files, plus a translation editing tool for .sdlxliff files. For the AI back end, the README states the app works with more than 40 providers, naming OpenAI, Google Gemini, Anthropic Claude, DeepSeek, xAI Grok, and Mistral, as well as local models through Ollama. It adds that any OpenAI compatible model or endpoint also works. This breadth means a user can pick a cloud service, a local model, or a mix, and the same interface drives all of them. The profile system for Korean, Japanese, and Chinese lets the tool apply language specific settings without manual setup each time. The README frames provider choice as a core feature, so the user is not locked to one API. The many supported formats make the app a single place for several translation tasks. The count of more than 40 providers shows the app is built around pluggable back ends rather than a single service. The information above is drawn from the project README and its repository metadata, which remain the source for the project's current behaviour. The app is built with PySide6, requires Python 3.10 or newer, and is licensed under MIT. A reader who wants exact commands or file names should open the repository and read the documentation the maintainers provide.
The manga and editing panels
Beyond plain text, the README describes two specialised panels. The manga translator panel combines multiple AI models to translate .cbz archives and direct image files, which addresses the harder problem of text inside images rather than in a text layer. The translation editing tool works with .sdlxliff files, a format used by computer aided translation editors, so a user can refine machine output in a familiar workflow. These panels extend Glossarion past document translation into visual and professional editing territory. The README presents them as part of the same suite, so a user dealing with a manga and its subtitle file can stay in one application. The combination of models in the manga panel suggests the tool chains providers to improve accuracy on image text. By including an editing step, the app acknowledges that raw AI output often needs human correction, and it gives a place to do that without leaving the GUI. The .sdlxliff support is a notable detail because it connects Glossarion to standard translation tooling, meaning work can move between this app and other editors without format loss. The information above is drawn from the project README and its repository metadata, which remain the source for the project's current behaviour. The app is built with PySide6, requires Python 3.10 or newer, and is licensed under MIT. A reader who wants exact commands or file names should open the repository and read the documentation the maintainers provide.
Reading the user guide
The README points to a complete user guide at docs/Glossarion_User_Guide_Aug-21-2026.md. It calls the guide the place where every button and setting is explained in plain language, and it notes the guide includes how to translate for free. The README also shows build, license, and Python version badges, plus links to releases and a Discord community. For a new user, the guide is the first stop after install, since it covers both setup and daily use. The Discord link gives a place to ask questions beyond the written docs. The README's promise that the guide explains free translation options matters to users who do not want to pay for an API key right away. By shipping the guide inside the repository as a dated markdown file, the project keeps documentation next to the code and versioned with it, so the instructions match the release a user downloaded. The dated filename also makes it clear which guide belongs to which version of the application. The information above is drawn from the project README and its repository metadata, which remain the source for the project's current behaviour. The app is built with PySide6, requires Python 3.10 or newer, and is licensed under MIT. A reader who wants exact commands or file names should open the repository and read the documentation the maintainers provide.
Editorial conclusion
Glossarion is licensed under MIT, written in Python 3.10 or newer, and ships a user guide dated August 21, 2026.
Community notes